Stock Markets May 14, 2026 07:55 AM

Wolfe Reaffirms Nvidia as Its Top AI Chip Pick Despite Short-Term Underperformance

Research house raises estimates on key AI chip names and points to rising hyperscaler spending as a structural tailwind

By Sofia Navarro NVDA AVGO MRVL AMD

Wolfe Research has maintained a bullish posture on AI chip makers heading into earnings season, upgrading its estimates for Nvidia and Broadcom while again naming Nvidia its highest-conviction idea. Analyst Chris Caso said earlier cloud capex worries have eased after strong first-quarter reports and that hyperscaler spending intentions continue to climb, leaving major cloud providers little choice but to invest in AI compute.

Wolfe Reaffirms Nvidia as Its Top AI Chip Pick Despite Short-Term Underperformance
NVDA AVGO MRVL AMD

Key Points

  • Wolfe Research raised estimates for Nvidia and Broadcom and reiterated Outperform ratings on Nvidia, Broadcom and Marvell, naming Nvidia its top pick.
  • Analyst Chris Caso said cloud capex concerns have largely faded after strong first-quarter earnings, and hyperscaler spending estimates continue to rise, supporting demand for AI compute.
  • Nvidia has underperformed peers over the past six weeks despite a roughly 28% gain overall, with Wolfe attributing the relative weakness to limited visibility on calendar-year 2027 revenue.

Wolfe Research reiterated a pro-growth stance on artificial intelligence semiconductor names ahead of upcoming earnings, increasing estimates for Nvidia and Broadcom and reaffirming Nvidia as its primary pick, even as the stock has lagged some peers over recent weeks.

Analyst Chris Caso highlighted that the capital expenditure concerns in the cloud sector that pressured AI compute stocks earlier in the year have largely dissipated following robust first-quarter earnings from major cloud players. According to Wolfe, hyperscaler spending estimates are continuing to rise, providing a clearer demand backdrop for AI hardware vendors.


Wolfe framed the wider investment dynamic around the emergence of agentic AI, arguing it leaves leading cloud providers with little latitude to reduce investment. As the firm put it, "hyperscalers simply have no choice but to spend," noting that foregoing investment would mean stepping away from a major technological shift.

Despite gaining roughly 28% over a broader timeframe, Nvidia has underperformed its AI compute peers over the past six weeks. Competitors such as Broadcom, Marvell Technology and Advanced Micro Devices have produced stronger short-term price runs during that period. Wolfe attributes Nvidia's relative weakness primarily to a lack of visibility into calendar-year 2027 revenue, a detail some competitors have addressed more explicitly with forward guidance.

The research team also pointed to Nvidia's $1 trillion disclosure at its GTC conference, saying that figure did not fully capture the company's total opportunity set. Wolfe noted additional sources of revenue remain to be booked and highlighted potential contributions from the Rubin pod architecture. The firm suggested that clearer guidance for 2027 could help Nvidia close the performance gap with its peers.

Wolfe acknowledged certain headwinds cited by the market - specifically potential share losses to custom chips and margin pressure stemming from memory pricing - but indicated these risks do not fundamentally change its longer-term outlook for Nvidia and other AI chip suppliers.


Wolfe Research reiterated Outperform ratings on Nvidia, Broadcom and Marvell, and emphasized that "NVDA remains our best idea. The stock's underperformance hasn't changed our fundamental view." The firm's comments signal continued conviction in AI compute demand driven by hyperscaler investment plans and the broader rollout of agentic AI capabilities.

What this means for markets and sectors: Wolfe's analysis centers on the interplay between hyperscaler capital spending and semiconductor vendor revenue trajectories, with implications for cloud infrastructure, AI compute suppliers, and the broader semiconductor sector.

Risks

  • Lack of clear calendar-year 2027 revenue guidance for Nvidia could keep the stock trading below peers - this affects valuations in the semiconductor and AI compute sectors.
  • Potential share losses to custom chips represent a competitive risk for established AI chipmakers and could alter market share dynamics within the semiconductor industry.
  • Margin pressure from memory pricing could compress profitability for AI hardware vendors and influence earnings in the semiconductor sector.

More from Stock Markets

INmune Bio Rallies After FDA Grants Fast Track Status to XPro for Early Alzheimer’s May 14, 2026 Gilt Yields Slide to Three-Day Low After Health Minister Resignation Announcement May 14, 2026 Insiders Buy Stabilis Solutions Stock; Shares Tick Up in Premarket Trading May 14, 2026 Baltic Dry Index Climbs to Five-Month Peak as Panamax and Supramax Support Gains May 14, 2026 Tadawul slips as energy, construction and cement lead losses May 14, 2026